Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Geophysics*R*&*D*needed*to*support*monitoring*for*compliance*with*test9ban*trea;es*(*LTBT,*NPT,*TTBT,*
PNET,*and*CTBT)*
Paul*G.*Richards*Lamont9Doherty*Earth*Observatory*
Columbia*University*
The four test ban treaties:
LTBT = Limited Test Ban Treaty (1963)
also known as the Partial Test Ban Treaty orthe Atmospheric Test Ban Treaty
trilateral (USA, USSR, UK; became multilateral)
prohibits nuclear weapons tests “or any othernuclear explosion” in the atmosphere, in outerspace, and under water.
Allowed underground nuclear testing (> 1500since 1963 – promoted monitoring capability).
TTBT = Threshold Test Ban Treaty (1974)
bilateral (USA, USSR)
banned underground nuclear weapons tests ofyield greater than 150 kilotons after March 1976
not ratified until 1990 (!) after an extensive seriesof battles over technical issues (yield estimation)
PNET = Peaceful Nuclear Explosions Treaty
bilateral (USA, USSR)
in effect banned underground nuclear explosionsdone for non-military purposes above 150 kt
(building underground cavities,putting out oil-well firesseismic sources for geophysical surveysmaking transuranic elementsconstruction of dams, canals...)
NPT = Non-Proliferation Treaty (1970)
the most important nuclear arms control treaty
multilateralinternational monitoring provided by the IAEA
bans transfer of nuclear weapons technologybetween non-nuclear and nuclear weapon states
had a section (Article V) stating that:“potential benefits from any peaceful applications of nuclearexplosions will be made available to non-nuclear-weapon StatesParty to the Treaty on a nondiscriminatory basis and ... the charge tosuch Parties for the explosive devices used will be as low as possibleand exclude any charge for research and development”
Treaty on Principles Governing the Activities ofStates in the Exploration and Use of OuterSpace, including the Moon and Other CelestialBodies
[negotiated and in effect, 1967]
includes:Article IVStates Parties to the Treaty undertake not to place in orbit around theEarth any objects carrying nuclear weapons or any other kinds ofweapons of mass destruction, install such weapons on celestial bodies, orstation such weapons in outer space in any other manner.
like the LTBT: not associated with formal methods of verification
But these are all now largelysuperceded by the...
CTBT = Comprehensive Nuclear-Test-Ban Treaty
negotiated from 1958 to 1996, though thistreaty has still not entered into force (why not?),so in effect we have a nuclear testing moratoriumat least by the recognized nuclear weapons states
Has the most extensively-developed verificationprocedures of any modern nuclear arms controltreaty (six global networks, big budget...)
(on-site verification provisions, similar in somerespects to the Chemical Weapons Convention)
Contributions of key technologies to CTBT monitoring of different test environments
KeyTechnologies
Underground Underwater Atmosphere Near Space
Seismic* major major secondary none
Radionuclide* major major major none
Hydroacoustic* secondary major secondary none
Infrasound* secondary secondary major none
Electromagnetic secondary secondary major major
Satellite Imagery major major secondary secondary
* technologies used by the International Monitoring System (Vienna)
Environmentof test
Six different steps in nuclear explosion monitoring:Detection
(did a particular station detect a useful signal?) Association
(can we gather all the different signals from the same “event”?) Location
(where was it?) Identification
(was it an earthquake, a mining blast, a nuclear weapon test?) Attribution
(if it was a nuclear test, what country carried it out?) Yield estimation
(how big was it?)
The CTBT will in practice be monitored by:
• the international CTBT Organization in Vienna, Austria;
• National Technical Means, which for the United States includesthe Atomic Energy Detection System (AEDS) operated by theAir Force Technical Applications Center (AFTAC); and
• the loosely organized efforts of numerous institutions, acquiringand processing data originally recorded for purposes other thantreaty monitoring
Hundreds of institutions continuously operate thousands ofseismometers.Seismically active regions of North America, Europe, Asia, Northand South Africa, and the Middle East are now routinely monitoreddown to low magnitudes in order to evaluate earthquake hazards.
The main idea.:
use archives of seismic events as an aid tohelp with monitoring events occurring today.
First, briefly review the main results of a study of Kazakhstanseismicity, using cross-correlation to detect small events:
Multi-Station Validation of Waveform Correlation Techniquesas applied to Broad Regional MonitoringMegan Slinkard, David Schaff, Natalya Mikhailova, Stephen Heck,Christopher Young, Paul G. Richards
(just accepted for publication in the Bulletin of theSeismological Society of America).
First, briefly review the main results of a study of Kazakhstanseismicity, using cross-correlation to detect small events:
Multi-Station Validation of Waveform Correlation Techniquesas applied to Broad Regional MonitoringMegan Slinkard, David Schaff, Natalya Mikhailova, Stephen Heck,Christopher Young, Paul G. Richards
(just accepted for publication in the Bulletin of theSeismological Society of America).
Second, describe preliminary results of cross-correlation as adetector applied to China seismicity, combined with cross-correlation to measure relative arrival times, to enable preciseestimates of event location.
The LEB catalog had 8015 origins in a 3 year period
MOTIVATION AND OBJECTIVES
Waveform correlation techniques have garnered increasing attention in the last few years, as their value in detecting and classifying repeated events has been demonstrated again and again[1]. In this research, we show the potential in extending waveform correlation techniques to broad regional monitoring for the bene�it of nuclear monitoring. We use data from the CTBTO’s International Monitoring System, which consists of a sparse network of stations that must monitor the entire globe.
INTRODUCTION
The KNDC regional catalog had over 45000, AFTER mining events were removed (except for mining events in the two boxed regions in Russia)
Comparing the CTBTO’s LEB catalog to a regional catalog from the Kazakhstan National Data Center (KNDC), which covers central Asia, we note the potential for waveform correlation to enhance the completeness of the LEB catalog.
BVAR BVAR
KURK KURK
!!!
A total of 1938 events generated the template waveforms used at the three stations. This Venn diagram shows how many events had
templates at one, or at two, or at all three stations.
!Station!
!Number!of!Detections!
!Detections!confirmed!in!the!LEB!
!!!!
!Detections!confirmed!in!the!KNDC!Catalog!
!Detections!confirmed!in!either!catalog!
!Detections!confirmed!by!
high!correlation!(>0.7)!
Detections!confirmed!by!!2!stations!
!(1481!seen!by!all!3!stations)!
!Confirmed!in!some!manner!
!MKAR!
!(array)!
!7426!
!!(from!506!of!811!templates)!
!526!
!!!!
!553!
!927!
!1309!
!!(183!not!in!LEB!or!
confirmed!by!another!station)!
!4740!(64%)!
!(4226!not!in!
LEB)!
!5136!(69%)!
!!(4610!not!in!
LEB)!
!BVAR!
!(array)!
!3096!
!(from!193!of!543!templates)!
!154!!!
!67!
!180!
!383!!
(17!not!in!LEB!or!confirmed!by!another!station)!
!2199!!(71%)!
!(2054!not!in!
LEB)!
!2307!(75%)!
!(2153!not!in!
LEB)!
!KURK!
!(3C)!
!!
!8526!
!(from!837!of!
1515!templates)!
!693!!!
!705!
!1179!
!2031!
!(376!not!in!LEB!or!
confirmed!by!another!station)!
!4895!(57%)!
!(4327!not!in!
LEB)!
!5648!(66%)!
!(4955!not!in!
LEB)!
!!!! !
!Station!
!Number!of!Detections!
!Detections!confirmed!in!the!LEB!
!!!!
!Detections!confirmed!in!the!KNDC!Catalog!
!Detections!confirmed!in!either!catalog!
!Detections!confirmed!by!
high!correlation!(>0.7)!
Detections!confirmed!by!!2!stations!
!(1481!seen!by!all!3!stations)!
!
Confirmed!in!some!manner!
!MKAR!
!(array)!
!7426!
!!(from!506!of!811!templates)!
!526!
!!!!
!553!
!927!
!1309!
!!(183!not!in!LEB!or!
confirmed!by!another!station)!
!
4740!(64%)!
!(4226!not!in!
LEB)!
!
5136!(69%)!
!!(4610!not!in!
LEB)!
!BVAR!
!(array)!
!3096!
!(from!193!of!543!templates)!
!154!!!
!67!
!180!
!383!!
(17!not!in!LEB!or!confirmed!by!another!station)!
!2199!!(71%)!
!(2054!not!in!
LEB)!
!
2307!(75%)!
!(2153!not!in!
LEB)!
!KURK!
!(3C)!
!!
!8526!
!(from!837!of!
1515!templates)!
!693!!!
!705!
!1179!
!2031!
!(376!not!in!LEB!or!
confirmed!by!another!station)!
!
4895!(57%)!
!(4327!not!in!
LEB)!
!
5648!(66%)!
!(4955!not!in!
LEB)!
!Summary:!templates!were!used!to!detect!more!than!ten!times!as!many!smaller!events!in!continuous!data!over!three!years,!and!most!of!them!were!confirmed.!
!
Total number of template waveforms, derived from three years,2006 to 2008, and applied to the same three year period:
MKAR (811 events, 9 channels) 7299 waveforms
BVAR (543 events, 9 channels) 4887 waveforms
KURK (1515 events, 3 channels) 4545 waveforms.
Data filtered to pass 0.5 – 5 hz; 25 second windows (Lg wave).
We searched all 21 channels using a 32 processor network.
We used a detection threshold set separately for each template,for a false alarm rate of about one wrong detection per year.
Total number of template waveforms, derived from three years,2006 to 2008, and applied to the same three year period:
MKAR (811 events, 9 channels) 7299 waveforms
BVAR (543 events, 9 channels) 4887 waveforms
KURK (1515 events, 3 channels) 4545 waveforms.
Data filtered to pass 0.5 – 5 hz; 25 second windows (Lg wave).
We searched all 21 channels using a 32 processor network.
We used a detection threshold set separately for each template,for a false alarm rate of about one wrong detection per year.
This dataset required only 2.5 days for the computation.
−8 −6 −4 −2 0 2 4 6 8 10
−4
−3
−2
−1
0
1
2
3
4
5
east (km)
north
(km
)
10 events; mean semi-axes of 95% error ellipses are a = 0.72 km and b = 1.15 km
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
east (km)
north
(km
)
9 events; mean semi-axes of 95% error ellipses are a = 0.03 km and b = 0.04 km
−1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1
−0.5
−0.4
−0.3
−0.2
−0.1
0
0.1
0.2
0.3
0.4
0.5
east (km)
north
(km
)
9 events; mean semi-axes of 95% error ellipses are a = 0.03 km and b = 0.04 km tens of meters!
!
North Korean Nuclear Test Site & Seismographic Stations in the Region
124˚E 125˚E 126˚E 127˚E 128˚E 129˚E 130˚E 131˚E 132˚E37˚N
38˚N
39˚N
40˚N
41˚N
42˚N
43˚N
44˚N
45˚N
0 100 200 300
km
2006
2009
2013Kilju
KSRS
USRK
INCN
MDJ
YABA
DHAA
CBAI
THAA
1,5
2
3
4
6
7
8
9
10
CN2
VLA
200 km
Yellow
Sea
East Sea
Japan Sea
North Korea
South Korea
China
Russia
Kaesong
WonsanPyongyang
Seoul
SinuijuHamhung
Chongjin
Kimchaek
Youngbyun
Nuclear testsEarthquakesKnown explosionsGSN stationsIMS stationsNational stations
120˚E 130˚E 140˚E
40˚N
50˚N
500 km
Russ
ia
Japan
China
KoreaEast Sea
Japan SeaINCN
MDJ
MAJO
BJT
HIA
KSRS
USRK
HIA
MJAR
Science & Global Security, 20:1–29, 2012Copyright C⃝ FOIISSN: 0892-9882 print / 1547-7800 onlineDOI: 10.1080/08929882.2012.652558
Radionuclide Evidence forLow-Yield Nuclear Testing inNorth Korea in April/May 2010
Lars-Erik De GeerSwedish Defense Research Agency, Stockholm, Sweden
Between 13 and 23 May 2010, four atmospheric radionuclide surveillance stations, inSouth Korea, Japan, and the Russian Federation, detected xenon and xenon daughterradionuclides in concentrations up to 10 and 0.1 mBq/m3 respectively. All these mea-surements were made in air masses that had passed over North Korea a few daysearlier. This article shows that these radionuclide observations are consistent with aNorth Korean low-yield nuclear test on 11 May 2010, even though no seismic signalsfrom such a test have been detected. Appendix 1 presents a detailed analysis of theradioxenon data and Appendix 2 describes a hypothetical nuclear test scenario consis-tent with this analysis, including the possibility that the test used uranium-235 ratherthan plutonium-239. The analysis suggests that the technical and analytical basis todetect small nuclear tests using radionuclide signatures may be more developed thanis generally assumed.
INTRODUCTION
North Korea conducted its first nuclear test explosion on 9 October 2006. Thetest was carried out underground in a deep tunnel1 and had an estimated yieldof approximately 0.9 kt.2 Due to the low yield its nuclear character was firstquestioned, but it was soon confirmed nuclear by regional and distant detec-tions of mBq/m3 range radioactive xenon isotopes.3,4 Then, on 25 May 2009, a
Received 20 May 2011; accepted 20 October 2011.Address correspondence to Lars-Erik De Geer, FOI, Swedish Defense Research Agency,SE-164 90, Stockholm, Sweden. E-mail: [email protected] views expressed in this publication are those of the author and do not neces-sarily reflect the views of the Swedish Defense Research Agency, the ComprehensiveNuclear-Test-Ban Treaty Organization, the South Korean Government, or any of theinstitutions involved in the routine reporting of data to the International Data Centre.Data made available from these sources is, however, greatly appreciated. The experthelp in applying the WebGrape software to a non-CTBT station that was provided byDr. Gerhard Wotawa at the Zentralanstalt fur Meteorologie und Geodynamik in Vienna,Austria, is also greatly appreciated.
1
Science & Global Security, 20:155–171, 2012Copyright C⃝ Taylor & Francis Group, LLCISSN: 0892-9882 print / 1547-7800 onlineDOI: 10.1080/08929882.2012.711183
Seismological Constraintson Proposed Low-YieldNuclear Testing in ParticularRegions and Time Periodsin the Past, with Commentson “Radionuclide Evidencefor Low-Yield Nuclear Testingin North Korea in April/May2010” by Lars-Erik De Geer
David P. Schaff, Won-Young Kim, and Paul G. RichardsLamont-Doherty Earth Observatory of Columbia University, Palisades, NY, USA
We have attempted to detect seismic signals from small explosions in North Korea onfive specific days in 2010 that feature in scenarios proposed by De Geer. We searchedthe seismic data recorded by station MDJ in northeastern China, applying three-component cross-correlation methods using signals from known explosions as tem-plates. We assess the capability of this method of detection, and of simpler methods,all of which failed to find seismic signals that would be expected if De Geer’s scenarioswere valid. We conclude that no well-coupled underground explosion above about a tonoccurred near the North Korea test site on these five days and that any explosion wouldhave to be very small (local magnitude less than about 2) to escape detection.
Received 29 April 2012; accepted 3 July 2012.This work was partially supported by the Defense Threat Reduction Agency underaward # HDTRA1-11-1-0027. This article is Lamont-Doherty Earth Observatory con-tribution number 7583.Address correspondence to Paul G. Richards, Lamont-Doherty Earth Observatory ofColumbia University, 61 Route 9W, PO Box 1000, Palisades, NY 10964, USA. E-mail:[email protected]
155
Projects planned under the CVT, at Lamont
( )include:
• explain the basis for a classical discriminant(mb – Ms), and why it works so well
• explain why it nearly failed for the 3 nucleartests by North Korea (2006, 2009, 2013)[but note that other methods worked well]
• improve estimates of the depth of a seismicsource (whether earthquake or explosion)
• produce a “Glasstone and Dolan” type ofreview of the technical aspects of nuclearexplosion monitoring – for various audiences
[hundreds of “grey literature” papers][“various audiences” – not just undergrads
and grad students ...]